Mailing List archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[vdr] Re: Unkown Picture Type 6



On Thu, Nov 06, 2003 at 11:58:21AM +0100, Peter Funk wrote:
> Hello,
> 
> Michael Möhle wrote:
> > > Do you think its a interrupt-Problem ?
> > 
> > You said that you changed your BS. Newer SUSEs use ACPI. Maybe there
> > was a change in the IRQs because of that.
> 
> To provide just another data point:
> On my machine the interupt assignment didn't change due to the system
> upgrade from SuSE 7.1/Kernel 2.2.19/vdr 1.0.4 to the c't VDR Distro
> I'm using right now.  
> 
> I've saa7146(0) on IRQ 11 and saa7146(1) on IRQ 9.  Both interupts are 
> exclusively assigned to the DVB-S Rev. 1.3 cards.  No IRQ sharing.  
> 
> Since rebooting fixes the problem for a while, I also can't believe, that 
> the problem is caused by a hardware failure in one of my cards.  It is
> very unlikely that one of the cards, which used to work flawlessly since
> February 2001 broke just at the moment, when I upgraded the VDR software.  
> Since other people reported the same kind of problems here, this seems
> to be even more unlikely.

[...]

> 
> So at the moment the problem occurs on my machine a daily basis.  
> This is annoying.  (Please don't misunderstand me: I'm still very 
> thankful about what Klaus and the Metzler Bros.
> and all you other guys have accomplished here).
> 
> Nevertheless I don't want to go back to VDR 1.0.4.  
> 
> So I would like to know, if I better should first try to change from
> the http://heise.de source/ vdr 1.2.2-5woody1 package distributed with
> the c't VDR distro to the recently released 1.2.5 or 1.2.6 or whether
> I should first starting to build an unstripped binary of vdr 1.2.2
> and try to attach with gdb to the VDR thread, whenever the problem
> happens next time.  But what to do then?  Where to look?  If it
> is problem within the firmware or driver, debugging in VDR wouldn't
> help to gather any useful information.
> 
> I've had a look at the vdr-1.2.5_Old_CA_Descriptorhandling.diff
> patch and I'm reluctant to apply it without understanding anything
> what it does.  As has already been discussed here, it seems to change 
> far to much at once.

Just a guess: maybe it would help to slow down the PCI access
a bit. If you found a BIOS entry like

        PCI Delay Transaction

and

        PCI#2 Access #1 Retry

you should enable both.  Also decreasing the PCI version may help
(hey DVB cards are old hardware).


        Werner


-- 
Info:
To unsubscribe send a mail to ecartis@linuxtv.org with "unsubscribe vdr" as subject.



Home | Main Index | Thread Index